Sediment volume partitioning: rationale for stratigraphic model evaluation and high-resolution stratigraphic correlation

Quantitative stratigraphic forward models attempt to simulate realistic stratigraphic attributes using geologically reasonable values of process parameters. Models assume different behaviors of the stratigraphic process-response system and employ different algorithms and mathematical simplifications to achieve the simulations. How does one judge whether a stratigraphic forward model produces geologically realistic simulations? We propose six criteria to evaluate strengths and Limits of applicability of stratigraphic models. Although the criteria are universal, not all models attempt to simulate all of them: only criteria appropriate to the particular model simulations need be used in the evaluation We consider a model robust and potentially accurate if it simulates: (1) sediment volume partitioning into different facies tracts (area in 2-D simulations is a proxy for volume in 3-D simulations): (2) changes in cycle symmetry through time and space: (3) an inverse distribution of facies tract heterogeneity and frequency of stratigraphic discontinuity surfaces; and (4) reasonable angles, geometries and distances of clinoforms. Like other numerical models of physical systems, stratigraphic models also must (5) obey applicable conservation Laws, and (6) produce simulations that do not change appreciably with small changes in temporal sample rates. A match with these criteria should be achieved using process parameter values that are geologically reasonable, and that are not derived directly from the data sets used for comparison with model outputs. As an example, we present a two-dimensional stratigraphic forward model for siliciclastic sediment accumulation in coastal plain through shelf environments. Input parameters are eustasy, tectonics, flexural strength of lithosphere, sediment supply. and initial depositional topography. Erosion, transfer and deposition of sediment are simulated by a calculated geometric approximation of stratigraphic base level. This approximation changes shape dynamically during a model run as a function of processes that modulate accommodation space, sediment supply, and depositional topography. The model satisfies all criteria other than criterion (3) which is not applicable for evaluation because the model does not simulate sedimentary facies. Model simulations illustrate sediment volume partitioning along with geographic and stratigraphic changes in cycle symmetries and frequency of surfaces of stratigraphic discontinuity. These simulations show the conceptual basis for high-resolution stratigraphic correlation and construction of rigorous correlation procedures for nonmarine to marine shelf strata. This rationale for correlation is based on a philosophy that considers the stratigraphic record complete - that time is completely represented in the form of rocks plus surfaces of stratigraphic discontinuity. Because surfaces of stratigraphic discontinuity form contemporaneously with sediment accumulation, but at different geographic positions, the simulations illustrate some of the geologic conditions for correlating rocks to rocks, rocks to surfaces, or surfaces to surfaces.